In this week's Stansberry Investor Hour, Dan welcomes Marko Papic back to the show. Marko is the chief strategist and head of GeoMacro at BCA Research, a global investment research firm.

Marko kicks things off by discussing the "second derivative of AI capex," which signals the beginning of the end of the AI boom. Due to tension in the Middle East potentially starting to ease up, the market is nearing the peak of the "Wall of Worry," and as a result, investors could lose a component that helps fuel the current rally. Additionally, Marko says that AI is inflationary. It takes labor, copper, and electricity to construct and run a data center, and with oil prices not likely to return to the levels they were at before the conflict at the Strait of Hormuz, that will just compound the inflation. Marko details what you can expect from the "endgame" of the Hormuz blockade. (0:00)

Next, Marko delves into oil prices and demand. He says that the conflict is starting to give several impressions to other countries after this passes. The first is that the U.S. creates demand when it has a desire to obtain resources and seeks them out. Countries will then start hoarding them as a means of securing them. The second impression the conflict shows is that our allies might not be able to rely on us in a prolonged conflict. Marko says that the raid in Venezuela earlier this year and the Strait of Hormuz situation were both supposed to be short-term incidents. The U.S. did not intend for the blockade to last as long as it has. So in the event of a drawn-out conflict, our allies might have second thoughts about asking for aid. However, even if we are shut out, Marko says America is integrated into the global infrastructure. (18:24)

Finally, Marko sums up the three main reasons why an "inflationary brew" is developing for data centers. The first is that Federal Reserve Chair Kevin Warsh might not be as dovish as hoped prior to entering the role. And it doesn't seem like President Donald Trump will do much to deter him from raising interest rates. That will make building data centers more expensive. The second is that the major AI IPOs are creating a massive supply with little liquidity. With many individual investors primarily having exposure to the S&P 500 Index, they'll be gaining exposure with their 401(k)s but won't be actively buying or selling them, resulting in stagnancy. And lastly, AI capex is slowing down since it's not feasible to build as many data centers as these companies desire. (36:06)